Applications of situation-dependent intensity and intensity spread predictions based on a weighted analog technique

A version of our situation-dependent intensity prediction (SDIP) is proposed for operational application after three modifications: (i) Ten historical track analogs are matched with Joint Typhoon Warning Center (JTWC) official track forecasts rather than besttracks; (ii) Giving two times as much weight to the 72 h — 120 h portion of the track as to the 0–72 h portion to give higher rankings for analog tracks with similar landfall or recurvature positions and timing; and (iii) Weighting both the intensity prediction technique and a new intensity spread guidance product according to new rankings of the track analogs rather than assuming all track analogs are equally likely. These special matchings and weightings of the track analogs in this weighted-analog intensity (WANI) add skill in the 72–120 h forecast intervals in regions where landfalls occur. Viability as an operational technique is demonstrated as the WANI has only 1 kt larger mean absolute errors than the JTWC intensity errors from 12 h through 72 h, and the WANI is 5 kt (20%) better at 120 h. The WANI rank-weighted intensity spreads each 12 h among the 10 best historical track analogs are processed to reduce any intensity bias and calibrated to reduce (increase) the over-determined (under-determined) intensity spreads at early (later) forecast intervals. Thus, the situation-dependent intensity spread guidance is generated that will include about 68% of the verifying intensities at all forecast intervals. Four examples of the WANI intensity predictions and intensity spread guidance are presented to illustrate how the forecaster might use this information in potential landfall and intensity bifurcation situations.     

Evaluation of long-term trends in tropical cyclone intensity forecasts

Summary The National Hurricane Center and Joint Typhoon Warning Center operational tropical cyclone intensity forecasts for the three major northern hemisphere tropical cyclone basins (Atlantic, eastern North Pacific, and western North Pacific) for the past two decades are examined for long-term trends. Results show that there has been some marginal improvement in the mean absolute error at 24 and 48 h for the Atlantic and at 72 h for the east and west Pacific. A new metric that measures the percent variance of the observed intensity changes that is reduced by the forecast (variance reduction, VR) is defined to help account for inter-annual variability in forecast difficulty. Results show that there have been significant improvements in the VR of the official forecasts in the Atlantic, and some marginal improvement in the other two basins. The VR of the intensity guidance models was also examined. The improvement in the VR is due to the implementation of advanced statistical intensity prediction models and the operational version of the GFDL hurricane model in the mid-1990s. The skill of the operational intensity forecasts for the 5-year period ending in 2005 was determined by comparing the errors to those from simple statistical models with input from climatology and persistence. The intensity forecasts had significant skill out to 96 h in the Atlantic and out to 72 h in the east and west Pacific. The intensity forecasts are also compared to the operational track forecasts. The skill was comparable at 12 h, but the track forecasts were 2 to 5 times more skillful by 72 h. The track and intensity forecast error trends for the two-decade period were also compared. Results showed that the percentage track forecast improvement was almost an order of magnitude larger than that for intensity, indicating that intensity forecasting still has much room for improvement.     

Seven-Day Intensity and Intensity Spread Predictions in Bifurcation Situations with Guidance-On-Guidance for Western North Pacific Tropical Cyclones

An objective technique to detect and predict intensity bifurcation situations in a five-day Weighted Analog Intensity forecast technique for the western North Pacific (WAIP) has been extended to seven days. A hierarchical cluster analysis is applied to the N analog intensities to separate them into two clusters, which are considered to represent a substantial intensity bifurcation if a threshold maximum velocity difference of 15 kt is satisfied. Two important modifications have been made to develop the bifurcation version for seven-day WAIP forecasts. First, the number of track analogs has been increased from 10 analogs to 16 analogs, which results in larger sample sizes and better performance. Second, separate intensity bias corrections are calculated for the two cluster WAIP forecasts rather than using the same 16-analog intensity bias correction. If an always perfect selection of the correct cluster WAIP forecast of each bifurcation situation is made, a substantial improvement in the intensity mean absolute errors is achieved relative to the original WAIP forecasts based on all 16 of the best analogs. These perfect-cluster selection WAIP forecasts have smaller bias errors and are more highly correlated with the verifying intensities at all forecast intervals through 168 h. Furthermore, the Probability of Detection is improved for the perfect-cluster selection and more realistic intensity spreads are specified. A simple guidance-on-guidance technique is demonstrated to assist the forecasters in selecting the correct WAIP cluster forecast in bifurcation situations.     

Situation-dependent intensity skill metric and intensity spread guidance for western North Pacific tropical cyclones

A situation-dependent intensity prediction (SDIP) technique is developed for western North Pacific tropical cyclones that is based on the average of the intensity changes from the 10 best historical track analogs to the Joint Typhoon Warning Center best-tracks. The selection of the 10 best track analogs is also conditioned on the current intensity, and it is demonstrated that for a subsample of current intensities less than or equal to 35 kt the intensity mean absolute errors (MAEs) and biases are smaller than for the greater than 35 kt intensity subsample. The SDIP is demonstrated to have advantages as an intensity skill measure at forecast intervals beyond 36 h compared to the current climatology and persistence technique that uses only variables available at the initial time. The SDIP has significantly smaller intensity MAEs beyond 36 h with an almost 20% reduction at 120 h, has significantly smaller intensity biases than the present skill metric beyond 12 h, and explains 36% of the intensity variability at 120 h compared to 20% explained variance for the current technique. The probability distributions of intensities at 72 h and 120 h predicted by the SDIP are also a better match of the distribution of the verifying observations. Intensity spread guidance each 12 h to 120 h is developed from the intensity spread among the 10 best historical track analogs. The intensity spread is calibrated to ensure that the SDIP forecasts will have a probability of detection (PoD) of at least 68.26%. While this calibrated intensity spread is specifically for the SDIP technique, it would provide a first-order spread guidance for the PoD for the official intensity forecast, which would be useful intensity uncertainty information for forecasters and decision-makers.     

1999—2003年西北太平洋热带气旋综合预报的误差分析

应用1999—2003年中国中央气象台 (CMO)、日本气象厅 (JMA) 以及美国联合台风警报中心 (JTWC) 发布的西北太平洋热带气旋综合预报资料, 从总误差、逐年误差趋势、不同海区误差、不同路径趋势误差、不同强度趋势误差等5个方面对各预报中心的路径及强度预报结果进行分析, 结果表明:5年总的平均误差以JTWC的路径预报误差最小, 而JMA的强度预报较准确; 在不同海域, 各预报中心的路径预报能力各有优势, 但在热带气旋的强度预报方面, JMA的方法在各海区都较稳定; 对不同路径趋势热带气旋的预报方面, 除了南海转向热带气旋的路径预报比JMA和CMO稍差一些之外, JTWC的路径预报在大多数情况下都是好于或相当于JMA和CMO; 在不同强度变化趋势热带气旋的预报方面, JTWC在大多数情况下都优于其他中心。上述结果帮助业务和科技人员全面了解各预报中心的预报能力优劣, 也为今后改进我国的热带气旋预报提供有益的参考。     

Improved tropical cyclone intensity and intensity spread prediction in bifurcation situations

Bifurcation or bi-modal tropical cyclone intensity forecasts may arise due to uncertainty in the timing of formation, timing and magnitude of rapid intensification periods, or track forecast uncertainty leading to landfall or non-landfall or leading to interaction with warm- or cold-ocean eddies. An objective technique is developed and tested to detect these intensity bifurcation situations in our weighted-analog intensity (WANI) forecasts that are based on the 10 best historical analogs to the Joint Typhoon Warning Center (JTWC) official track forecasts. About 19% of the overall sample of 1136 WANI forecasts in the western North Pacific during the 2010–2012 seasons met the criteria for a substantial intensity bifurcation situation. Using a hierarchical clustering technique, two clusters of the 10 best analogs are defined and separate WANI forecasts and intensity spreads are calculated for the two clusters. If an always perfect selection of the correct cluster WANI forecast of each bifurcation situation is made, a substantial improvement in the intensity mean absolute errors is achieved relative to the original WANI forecasts based on all 10 of the best analogs. These perfect-cluster selection WANI forecasts have smaller bias errors and are more highly correlated with the verifying intensities at all forecast intervals through 120 h. Without further bias correction and calibration, the cluster WANI intensity spreads are under-determined as the Probability of Detections are smaller than the desired 68%. Four examples of WANI cluster predictions of intensity bifurcation situations are provided to illustrate how a correct choice of the intensity forecast and the intensity spread can be the basis for improved warnings of the threat from western North Pacific tropical cyclones.     

西北太平洋(含南海)热带气旋路径集成预报分析

基于2004—2009 年中国中央气象台、日本气象厅、美国联合台风警报中心、欧洲中心对西北太平洋和南海编号热带气旋主客观预报资料,利用算术平均、多元回归以及历史平均误差等三种集成方法,建立了热带气旋路径集成预报业务化系统。通过2007—2009 年的业务运行结果分析发现,欧洲中心客观预报参与的24、48 和72 h 集成比主观预报三个成员集成预报水平分别提高约2%、3%～5%和3%～5%,减小误差2.5 km左右、6～9 km 和10～12 km。技巧分析发现,24～72 h 集成预报有正技巧,多元回归集成技巧相对稍低,而算术平均和以各成员平均误差的平方倒数为权重系数的集成技巧对于各集成成员来说技巧差异不大。96 h 集成预报对欧洲中心的客观预报没有正技巧。      

OPERATIONAL ENSEMBLE FORECASTING AND ANALYSIS OF TROPICAL CYCLONES OVER THE WESTERN NORTH PACIFIC (INCLUDING THE SOUTH CHINA SEA)

Based on the tropical cyclone data from the Central Meteorological Observatory of China, Japan Meteorological Agency, Joint Typhoon Warning Center and European Centre for Medium-Range Weather Forecasts （ECMWF） during the period of 2004 to 2009, three consensus methods are used in tropical cyclone （TC） track forecasts. Operational consensus results show that the objective forecasts of ECMWF help to improve consensus skill by 2%, 3%-5% and 3%-5%, decrease track bias by 2.5 kin, 6-9 km and 10-12 km for the 24 h, 48 h and 72 h forecasts respectively over the years of 2007 to 2009. Analysis also indicates that consensus forecasts hold positive skills relative to each member. The multivariate regression composite is a method that shows relatively low skill, while the methods of arithmetic averaging and composite （in which the weighting coefficient is the reciprocal square of mean error of members） have almost comparable skills among members. Consensus forecast for a lead time of 96 h has negative skill relative to the ECMWF objective forecast.     

数值模式的热带气旋强度预报订正及其集成应用

提供热带气旋强度预报产品的业务数值天气预报模式有很多,并已表现出一定的预报技巧,为提高对模式热带气旋强度预报产品的定量应用能力,分析2010—2012年7个业务数值模式的西北太平洋热带气旋强度预报,发现预报误差不仅受到模式热带气旋初始强度误差的显著影响,还与热带气旋及其所处环境的初始状况有密切关系,包括热带气旋初始强度、尺度、移速、环境气压、环境风切变、热带气旋发展潜势等。根据这些因子与各模式热带气旋强度预报误差之间的相关性,采用逐步回归方法建立热带气旋强度预报误差的统计预估模型,并通过逐个热带气旋滚动式建模来进行独立样本检验。检验结果表明,基于误差预估的模式订正预报比模式直接输出的热带气旋强度预报有显著改进,在此基础上建立的热带气旋强度多模式集成预报方案相对气候持续性预报方法在12 h有28%的正技巧,在24—72 h则稳定在15%—20%,具有业务参考价值。     

INTENSITY CHANGE CHARACTERISTICS OF TROPICAL CYCLONES IN THE WESTERN NORTH PACIFIC AS REVEALED BY THREE DIFFERENT DATASETS

Analyzed in this paper are the 20-yr(1991-2010)tropical cyclone(TC)intensity from three forecast centers in the Western North Pacific,i.e.China Meteorological Administration(CMA),Japan Meteorological Agency(JMA),and Joint Typhoon Warning Center(JTWC)of the United States.Results show that there is more or less discrepancy in the intensity change of a TC among different datasets.The maximum discrepancy reaches 22 hPa/6h(42 hPa/6h,33 hPa/6h)between CMA and JMA(CMA and JTWC,JMA and JTWC).Special attention is paid to the records for abrupt intensity change,which is currently a difficult issue for forecasters globally.It is found that an abrupt intensity change process recorded by one dataset can have,in some extreme cases,intensity change in another dataset varying from 0 to≥10 hPa/6h with the same sign or the opposite sign.In a total of 2511 cases experiencing rapid intensity change,only 14%have consensus among all the three datasets and 25%have agreement between two of the three datasets.In spite of such a significant uncertainty,the three datasets agree on the general statistical characteristics of abrupt intensity change,including regional and seasonal distribution,the relationship with initial intensity and TC moving speed,and persistence features.Notable disagreement is on very strong systems(SuperTY)and TCs moving very fast.     

2009年夏季西太平洋台风路径和强度的多模式集成预报

基于TIGGE资料中的中国气象局、欧洲中期天气预报中心、日本气象厅和英国气象局等四个中心的2009年5月1日-8月31日台风预报资料,利用多模式集合平均、消除偏差集合平均和加权消除偏差集合平均等方法,对2009年8月1-31日预报期的西太平洋的台风路径和强度(中心气压)进行24～ 72 h预报时效的多模式集成预报,并对0907号台风“天鹅”和0908号台风“莫拉克”进行个例分析.结果表明:各中心对于不同时效的预报,预报技巧有明显差异.消除偏差集合平均与加权消除偏差集合平均显著地减小了预报误差,预报效果优于最好的单个中心预报和多模式集合平均.对于24 ～ 72 h预报,加权消除偏差集合平均方法始终表现出最好的预报性能.     

Experiments of DSAEF_LTP Model with Two Improved Parameters for Accumulated Precipitation of Landfalling Tropical Cyclones over Southeast China

The Dynamical-Statistical-Analog Ensemble Forecast model for landfalling tropical cyclones (TCs) precipitation (DSAEF_LTP) utilises an operational numerical weather prediction (NWP) model for the forecast track, while the precipitation forecast is obtained by finding analog cyclones, and making a precipitation forecast from an ensemble of the analogs. This study addresses TCs that occurred from 2004 to 2019 in Southeast China with 47 TCs as training samples and 18 TCs for independent forecast experiments. Experiments use four model versions. The control experiment DSAEF_LTP_1 includes three factors including TC track, landfall season, and TC intensity to determine analogs. Versions DSAEF_LTP_2, DSAEF_LTP_3, and DSAEF_LTP_4 respectively integrate improved similarity region, improved ensemble method, and improvements in both parameters. Results show that the DSAEF_LTP model with new values of similarity region and ensemble method (DSAEF_LTP_4) performs best in the simulation experiment, while the DSAEF_LTP model with new values only of ensemble method (DSAEF_LTP_3) performs best in the forecast experiment. The reason for the difference between simulation (training sample) and forecast (independent sample) may be that the proportion of TC with typical tracks (southeast to northwest movement or landfall over Southeast China) has changed significantly between samples. Forecast performance is compared with that of three global dynamical models (ECMWF, GRAPES, and GFS) and a regional dynamical model (SMS-WARMS). The DSAEF_LTP model performs better than the dynamical models and tends to produce more false alarms in accumulated forecast precipitation above 250 mm and 100 mm. Compared with TCs without heavy precipitation or typical tracks, TCs with these characteristics are better forecasted by the DSAEF_LTP model.     

中央气象台台风强度综合预报误差分析

本文从总误差、逐年趋势、误差分布等方面对2001—2012年中央气象台（Central Meteorological Observatory, CMO）的台风（TC）强度综合预报水平进行分析,初步分析了强度迅速变化台风预报偏差大的原因。结果表明,强度预报水平没有明显改善,预报误差呈现逐年波动状态,强度稳定TC的预报误差最小,迅速加强TC的预报误差最大。24、96~120 h预报偏强的概率较大,而48~72 h预报偏弱的概率大。南海东北部等海域的预报误差较大,应在业务预报中特别予以关注。随着TC强度的逐渐增强,强度预报在120 h内预报偏强的可能性变大,而强度预报偏弱的可能性减小。根据误差分析结果,提出了一个强度概率预报方案,检验结果表明可在业务中参考使用。     

2018年西北太平洋台风活动特征和预报难点分析

利用1949-2018年中国气象局台风最佳路径、2018年中央气象台的台风路径强度实时预报、ECMWF数值预报以及NCEP逐日高分辨率海温RTG_SST(0.083°×0.083°)等资料,对2018年西北太平洋台风活动的主要特征和预报难点进行了分析.结果 表明:2018年台风生成频数偏多,生成源地偏东,南海台风活跃;...     

对台风路径统计动力预报方案(SD-75)的统计参数作调整的试验

本文在台风路径统计动力预报方案(SD-75)扩大样本量以后,重新求出运动方程中的主要作用力的回归表达式。然后,利用调整后的方案(简称SD-85)对1987年台风路径的业务预报作了试验,结果表明该方案的业务预报性能优于SD-75预报方案。     

GRAPES-TCM业务试验结果分析

应用数值模式GRAPES-TCM2.1（GT2.1）的升级版GRAPES-TCM2.6（GT2.6）对2005年西北太平洋及南海热带气旋路径进行了后报试验和性能检验。在检验预报路径时考察了平均距离误差、系统偏差、距离误差的地理分布和平均移速误差,并在总体检验的基础上,根据与热带气旋路径相关的重要因素（路径类型、强度、有无登陆过程以及有无移向或移速突变）进行了分类检验。结果表明,GT2.6表现出良好的预报性能,其对所有样本的24、48和72小时平均距离误差分别为135.8、230.7和336.0km,但前12小时距离误差较大（近100km）。用系统偏差订正GT2.6对转向类样本的预报路径可获得明显的改进。GT2.6对130°E以西区域特别是靠近我国东南沿海的热带气旋路径预报表现较好。GT2.6的总体平均预报移速在48小时内都比最佳路径移速偏快,初始的12小时内偏快最多（近1m·s^-1）,在48小时后转为偏慢。GT2.6和GT2.1的前48小时路径预报性能的稳定性相当,总体误差特征相似,GT2.6对所有样本的48小时预报有显著改进,平均距离误差减小17km。初步分析表明,对GT2.6的改进应主要集中在数值预报初始场的改进方面,比如在初始化过程中采用较准确的涡旋分离方案、加入引导向量改进初始移动以及适当考虑加入非对称Bogus涡旋;扩大预报区域,并适当选取预报区域起始位置等。     

Possible Sources of Forecast Errors Generated by the Global/Regional Assimilation and Prediction System for Landfalling Tropical Cyclones. Part Ⅰ: Initial Uncertainties

This paper investigates the possible sources of errors associated with tropical cyclone(TC) tracks forecasted using the Global/Regional Assimilation and Prediction System(GRAPES). The GRAPES forecasts were made for 16 landfalling TCs in the western North Pacific basin during the 2008 and 2009 seasons, with a forecast length of 72 hours, and using the default initial conditions("initials", hereafter), which are from the NCEP-FNL dataset, as well as ECMWF initials. The forecasts are compared with ECMWF forecasts. The results show that in most TCs, the GRAPES forecasts are improved when using the ECMWF initials compared with the default initials. Compared with the ECMWF initials, the default initials produce lower intensity TCs and a lower intensity subtropical high, but a higher intensity South Asia high and monsoon trough, as well as a higher temperature but lower specific humidity at the TC center. Replacement of the geopotential height and wind fields with the ECMWF initials in and around the TC center at the initial time was found to be the most efficient way to improve the forecasts. In addition, TCs that showed the greatest improvement in forecast accuracy usually had the largest initial uncertainties in TC intensity and were usually in the intensifying phase. The results demonstrate the importance of the initial intensity for TC track forecasts made using GRAPES, and indicate the model is better in describing the intensifying phase than the decaying phase of TCs. Finally, the limit of the improvement indicates that the model error associated with GRAPES forecasts may be the main cause of poor forecasts of landfalling TCs. Thus, further examinations of the model errors are required.     

AN OPERATIONAL STATISTICAL SCHEME FOR TROPICAL CYCLONE INDUCED RAINFALL FORECAST

A non-parametric method is used in this study to analyze and predict short-term rainfall due to tropical cyclones(TCs) in a coastal meteorological station. All 427 TCs during 1953-2011 which made landfall along the Southeast China coast with a distance less than 700 km to a certain meteorological station- Shenzhen are analyzed and grouped according to their landfalling direction, distance and intensity. The corresponding daily rainfall records at Shenzhen Meteorological Station(SMS) during TCs landfalling period(a couple of days before and after TC landfall) are collected. The maximum daily rainfall(R-24) and maximum 3-day accumulative rainfall(R-72) records at SMS for each TC category are analyzed by a non-parametric statistical method, percentile estimation. The results are plotted by statistical boxplots, expressing in probability of precipitation. The performance of the statistical boxplots is evaluated to forecast the short-term rainfall at SMS during the TC seasons in 2012 and 2013. Results show that the boxplot scheme can be used as a valuable reference to predict the short-term rainfall at SMS due to TCs landfalling along the Southeast China coast.     

Possible sources of forecast errors generated by the global/regional assimilation and prediction system for landfalling tropical cyclones. Part I: Initial uncertainties

This paper investigates the possible sources of errors associated with tropical cyclone (TC) tracks forecasted using the Global/Regional Assimilation and Prediction System (GRAPES). The GRAPES forecasts were made for 16 landfalling TCs in the western North Pacific basin during the 2008 and 2009 seasons, with a forecast length of 72 hours, and using the default initial conditions (“initials”, hereafter), which are from the NCEP-FNL dataset, as well as ECMWF initials. The forecasts are compared with ECMWF forecasts. The results show that in most TCs, the GRAPES forecasts are improved when using the ECMWF initials compared with the default initials. Compared with the ECMWF initials, the default initials produce lower intensity TCs and a lower intensity subtropical high, but a higher intensity South Asia high and monsoon trough, as well as a higher temperature but lower specific humidity at the TC center. Replacement of the geopotential height and wind fields with the ECMWF initials in and around the TC center at the initial time was found to be the most efficient way to improve the forecasts. In addition, TCs that showed the greatest improvement in forecast accuracy usually had the largest initial uncertainties in TC intensity and were usually in the intensifying phase. The results demonstrate the importance of the initial intensity for TC track forecasts made using GRAPES, and indicate the model is better in describing the intensifying phase than the decaying phase of TCs. Finally, the limit of the improvement indicates that the model error associated with GRAPES forecasts may be the main cause of poor forecasts of landfalling TCs. Thus, further examinations of the model errors are required.     

近30a登陆我国的西北太平洋热带气旋活动的时空变化特征

采用1979—2006年美国联合台风预警中心的热带气旋（tropical cyclone,TC）资料,对登陆我国的西北太平洋（Northwest Pacific,NWP）TC强度、路径、登陆地点的气候特征、年际变化及其演变趋势进行了统计分析。结果表明：登陆我国的TC以发源于西北太平洋的西侧以及南海中、北部为主,并且在NWP西南区生成的登陆我国的TC基本以西北移动路径为主,而在NWP西北侧和南海生成的登陆我国的TC多为打转或移动路径转向;登陆我国的TC不仅在强度上具有明显增强的变化规律,而且在登陆位置上存在向东北方向偏移的演变趋势,使得登陆厦门以北区域的TC数量具有增加的趋势,而登陆厦门以南的TC数量存在减少的趋势;登陆我国的NWP TC移动路径存在年代际的演变特征。